Hi, I use samba 3.0.26a on fedora 8 as a fileserver for a win 2k3 domain. This has worked fine for about 2 months without any problems. However I came to the server 3 days ago and the harddrive was 100% full. On checking I found 60gb of core dumps in the winbind folder. I did a lot of searching and couldnt find anything relevent for this release. I tried upgrading samba to 3.0.28 (fc8 supplied rpm) and this does the same. The log.winbindd-idmap log suggests to me that it has a problem with ldap and empty results, so I made a quick script to check for gaps in the ldap records and found that several uid and gid numbers were not assigned (ie there was no entry for them in ldap, even though there were entries after them). Winbind does still mostly work just fails once in about every 10 tries. I believe it fails for both samba and dovecot (pop3/imap mail server). At the moment we are generating about 10gb/hour of core dumps which a cron job is keeping cleaned up. Has anyone got any ideas on this? also is it possible to tell samba/winbind not to do core dumps? smb.conf (only shown one share as we have lots!): [global] printer admin = @"MCS+sysadmin","MCS+root","MCS+administrator" add machine script = /usr/sbin/useradd -d /dev/null -g samba-clients -s /bin/false -M %u server string = Meadows log level = 1 syslog = 0 # vfs objects = extd_audit log file = /var/log/samba/%U.smbd.log max log size= 100000 printing = cups update encrypted = Yes encrypt passwords = Yes preferred master = no map to guest = Bad User name resolve order = lmhosts hosts wins bcast kernel oplocks = no oplocks = no locking = no level2 oplocks = no workgroup = MCS netbios name = MCS3 wins server = 192.168.0.8 wins support = no wins proxy = no admin users = "MCS+administrator" unix extensions = no security = ADS lanman auth = yes client lanman auth = yes client plaintext auth = yes idmap uid = 10000-100000 idmap gid = 10000-100000 ldap admin dn = cn=Manager,dc=meadows,dc=derbyshire,dc=sch,dc=uk ldap idmap suffix = ou=Idmap ldap suffix = dc=meadows,dc=derbyshire,dc=sch,dc=uk idmap backend = ldap:ldap://192.168.0.1 winbind use default domain = yes winbind separator = + winbind enum groups = yes winbind enum users = yes winbind cache time = 60 dns proxy = no password server = MCS1 template homedir = /home/%G/%U template shell = /bin/false realm = MEADOWS.DERBYSHIRE.SCH.UK [info] valid users = @"MCS+staff",@"MCS+teacher",@"MCS+sysadmin" write list = @"MCS+staff",@"MCS+teacher",@"MCS+sysadmin" path = /school/datadrive force directory mode = 0777 force create mode = 0666 comment = DataDrive create mode = 0666 directory mode = 777 browseable = yes log.winbindd-idmap: winbindd: ../../../libraries/libldap/getentry.c:48: ldap_next_entry: Assertion `entry != ((void *)0)' failed. [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(41) ==============================================================[2008/02/27 14:54:20, 0] lib/fault.c:fault_report(42) INTERNAL ERROR: Signal 6 in pid 29940 (3.0.28-0.fc8) Please read the Trouble-Shooting section of the Samba3-HOWTO [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(44) From: http://www.samba.org/samba/docs/Samba3-HOWTO.pdf [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(45) ==============================================================[2008/02/27 14:54:20, 0] lib/util.c:smb_panic(1655) PANIC (pid 29940): internal error [2008/02/27 14:54:20, 0] lib/util.c:log_stack_trace(1759) BACKTRACE: 19 stack frames: #0 winbindd(log_stack_trace+0x2d) [0xb7d515ad] #1 winbindd(smb_panic+0x5d) [0xb7d516dd] #2 winbindd [0xb7d3c10a] #3 [0x12d420] #4 [0x12d402] #5 /lib/libc.so.6(gsignal+0x50) [0x2f4690] #6 /lib/libc.so.6(abort+0x101) [0x2f5f91] #7 /lib/libc.so.6(__assert_fail+0xee) [0x2ed93e] #8 /usr/lib/libldap-2.3.so.0(ldap_next_entry+0x9c) [0x29d9fc] #9 winbindd [0xb7e8afa7] #10 winbindd(idmap_unixids_to_sids+0x397) [0xb7e84aa7] #11 winbindd(idmap_uid_to_sid+0x6b) [0xb7e888ab] #12 winbindd(winbindd_dual_uid2sid+0x61) [0xb7ce9fc1] #13 winbindd [0xb7ce7202] #14 winbindd [0xb7ce803f] #15 winbindd [0xb7cb9459] #16 winbindd(main+0x94d) [0xb7cb9e4d] #17 /lib/libc.so.6(__libc_start_main+0xe0) [0x2e1390] #18 winbindd [0xb7cb80f1] [2008/02/27 14:54:20, 0] lib/fault.c:dump_core(181) dumping core in /var/log/samba/cores/winbindd Thanks Robin This email has been processed by SmoothZap - www.smoothwall.net
Robin wrote:> Hi, > I use samba 3.0.26a on fedora 8 as a fileserver for a win 2k3 domain. This > has worked fine for about 2 months without any problems. However I came to > the server 3 days ago and the harddrive was 100% full. On checking I found > 60gb of core dumps in the winbind folder. I did a lot of searching and > couldnt find anything relevent for this release. I tried upgrading samba to > 3.0.28 (fc8 supplied rpm) and this does the same. The log.winbindd-idmap > log suggests to me that it has a problem with ldap and empty results, so I > made a quick script to check for gaps in the ldap records and found that > several uid and gid numbers were not assigned (ie there was no entry for > them in ldap, even though there were entries after them). > > Winbind does still mostly work just fails once in about every 10 tries. I > believe it fails for both samba and dovecot (pop3/imap mail server). At the > moment we are generating about 10gb/hour of core dumps which a cron job is > keeping cleaned up. Has anyone got any ideas on this? also is it possible > to tell samba/winbind not to do core dumps?enable core files = No Sorry, can't help with the ldap though. Regards, Doug
Robin, I have the exact same problem on one of our servers. Could you share the script you wrote, the one that checks for gaps in the ldap entries? Thanks, Pim> Hi, > I use samba 3.0.26a on fedora 8 as a fileserver for a win 2k3 domain. This > has worked fine for about 2 months without any problems. However I came to > the server 3 days ago and the harddrive was 100% full. On checking I found > 60gb of core dumps in the winbind folder. I did a lot of searching and > couldnt find anything relevent for this release. I tried upgrading samba to > 3.0.28 (fc8 supplied rpm) and this does the same. The log.winbindd-idmap > log suggests to me that it has a problem with ldap and empty results, so I > made a quick script to check for gaps in the ldap records and found that > several uid and gid numbers were not assigned (ie there was no entry for > them in ldap, even though there were entries after them). > > Winbind does still mostly work just fails once in about every 10 tries. I > believe it fails for both samba and dovecot (pop3/imap mail server). At the > moment we are generating about 10gb/hour of core dumps which a cron job is > keeping cleaned up. Has anyone got any ideas on this? also is it possible > to tell samba/winbind not to do core dumps? > > smb.conf (only shown one share as we have lots!): > > [global] > printer admin = @"MCS+sysadmin","MCS+root","MCS+administrator" > add machine script = /usr/sbin/useradd -d /dev/null -g samba-clients -s > /bin/false -M %u > server string = Meadows > log level = 1 > syslog = 0 > # vfs objects = extd_audit > log file = /var/log/samba/%U.smbd.log > max log size= 100000 > printing = cups > update encrypted = Yes > encrypt passwords = Yes > preferred master = no > map to guest = Bad User > name resolve order = lmhosts hosts wins bcast > kernel oplocks = no > oplocks = no > locking = no > level2 oplocks = no > workgroup = MCS > netbios name = MCS3 > wins server = 192.168.0.8 > wins support = no > wins proxy = no > admin users = "MCS+administrator" > unix extensions = no > > security = ADS > > lanman auth = yes > client lanman auth = yes > client plaintext auth = yes > idmap uid = 10000-100000 > idmap gid = 10000-100000 > ldap admin dn = cn=Manager,dc=meadows,dc=derbyshire,dc=sch,dc=uk > ldap idmap suffix = ou=Idmap > ldap suffix = dc=meadows,dc=derbyshire,dc=sch,dc=uk > idmap backend = ldap:ldap://192.168.0.1 > winbind use default domain = yes > winbind separator = + > winbind enum groups = yes > winbind enum users = yes > winbind cache time = 60 > dns proxy = no > password server = MCS1 > template homedir = /home/%G/%U > template shell = /bin/false > realm = MEADOWS.DERBYSHIRE.SCH.UK > > [info] > valid users = @"MCS+staff",@"MCS+teacher",@"MCS+sysadmin" > write list = @"MCS+staff",@"MCS+teacher",@"MCS+sysadmin" > path = /school/datadrive > force directory mode = 0777 > force create mode = 0666 > comment = DataDrive > create mode = 0666 > directory mode = 777 > browseable = yes > > log.winbindd-idmap: > > winbindd: ../../../libraries/libldap/getentry.c:48: ldap_next_entry: > Assertion `entry != ((void *)0)' failed. > [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(41) > ==============================================================> [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(42) > INTERNAL ERROR: Signal 6 in pid 29940 (3.0.28-0.fc8) > Please read the Trouble-Shooting section of the Samba3-HOWTO > [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(44) > > From: http://www.samba.org/samba/docs/Samba3-HOWTO.pdf > [2008/02/27 14:54:20, 0] lib/fault.c:fault_report(45) > ==============================================================> [2008/02/27 14:54:20, 0] lib/util.c:smb_panic(1655) > PANIC (pid 29940): internal error > [2008/02/27 14:54:20, 0] lib/util.c:log_stack_trace(1759) > BACKTRACE: 19 stack frames: > #0 winbindd(log_stack_trace+0x2d) [0xb7d515ad] > #1 winbindd(smb_panic+0x5d) [0xb7d516dd] > #2 winbindd [0xb7d3c10a] > #3 [0x12d420] > #4 [0x12d402] > #5 /lib/libc.so.6(gsignal+0x50) [0x2f4690] > #6 /lib/libc.so.6(abort+0x101) [0x2f5f91] > #7 /lib/libc.so.6(__assert_fail+0xee) [0x2ed93e] > #8 /usr/lib/libldap-2.3.so.0(ldap_next_entry+0x9c) [0x29d9fc] > #9 winbindd [0xb7e8afa7] > #10 winbindd(idmap_unixids_to_sids+0x397) [0xb7e84aa7] > #11 winbindd(idmap_uid_to_sid+0x6b) [0xb7e888ab] > #12 winbindd(winbindd_dual_uid2sid+0x61) [0xb7ce9fc1] > #13 winbindd [0xb7ce7202] > #14 winbindd [0xb7ce803f] > #15 winbindd [0xb7cb9459] > #16 winbindd(main+0x94d) [0xb7cb9e4d] > #17 /lib/libc.so.6(__libc_start_main+0xe0) [0x2e1390] > #18 winbindd [0xb7cb80f1] > [2008/02/27 14:54:20, 0] lib/fault.c:dump_core(181) > dumping core in /var/log/samba/cores/winbindd > > > Thanks > Robin > > > This email has been processed by SmoothZap - www.smoothwall.net > >